The number of protein structures has increased enormously in recent years, yet the ability to reliably predict the functional consequences of sequence changes remains elusive. This situation limits the ability to design drugs, predict drug metabolism, toxicity and resistance, make functional assignments from protein sequence, and construct novel enzymes de novo. Clearly a deeper understanding of how proteins with similar structures recognize different substrates and/or catalyze different chemical transformations is required. The ultimate manifestation of such understanding would be computational models that accurately predict function. Although remarkable advances have occurred in recent years, computational models of enzyme reactions are at best rudimentary. New computational methods are required. Progress demands the iterative deployment of new computational methods and detailed experimental characterization, which in turn requires close collaboration of computational and experimental scientists. This project proposes such a collaboration to develop computational models that accurately recapitulate and quantitatively predict the catalytic properties of proteins in the IMPDH/GMPR family. These enzymes have similar catalytic machinery, bind the same ligands with similar affinities, and react to form the same covalent intermediate, yet with very different outcomes. IMPDH catalyzes the conversion of inosine monophosphate (IMP) to xanthosine monophosphate (XMP) with the concomitant reduction of NAD+. This reaction controls the entry of purines into the guanine nucleotide pool, which in turn controls proliferation, making IMPDH a billion-dollar target for immunosuppressive and cancer chemotherapy. GMPR catalyzes the complimentary reduction of GMP to IMP and ammonia with concomitant oxidation of NADPH. The IMPDH reaction involves hydride transfer followed by hydrolysis while GMPR involves deamination followed by hydride transfer. How can a single active site accommodate multiple transition states? IMPDH solves this problem by rearranging its active site while GMPR uses two different cofactor conformations. Thus this project will require computational models of multiple chemical transformations with the accompanying conformational changes. The specific aims are: Aim 1. Develop a computational model for hydride transfer and test experimentally. Aim 2. Develop a computational model for the deamination reaction of GMPR and test experimentally.